Numbers to the right of topics indicate sequential lecture numbers.
Hn stands for Harrell Chapter n. Ln stands for lecture n.
Introduction (H1) L1
Course overview and logistics
Course philosophy
Hypothesis testing vs. estimation vs. prediction
Examples of multivariable prediction problems
Misunderstandings about classification vs. prediction
Study planning considerations L2
Choice of model
Model uncertainty/data driven model selection/phantom d.f.
General methods for multivariable models (H2) L3
Notation for general regression models
Model formulations
Interpreting model parameters
nominal predictors
interactions
Review of chunk tests
Relaxing linearity assumption for continuous predictors L4
avoiding categorization
nonparametric smoothing
simple nonlinear terms
splines for estimating shape of regression function and determining predictor transformations
cubic spline functions
restricted cubic splines
advantages of splines over other methods such as nonparametric regression
recursive partitioning and tree models in a nutshell
New directions in predictive modeling (L5)
Tests of association
Assessment of model fit
regression assumptions
modeling and testing complex interactions
distributional assumptions
Missing data (H3)
Types of missing data L6
Prelude to modeling
Problems with alternatives to imputation
Strategies for developing imputations
Single imputation
Multiple imputation
Predictive mean matching
The aregImpute algorithm
Multivariable modeling strategy (H4) L7
Pre-specification of predictor complexity
Variable selection
Overfitting and number of predictors
Shrinkage
Data reduction (H4.7, first page and summary chart, H14 up to H14.4, L8
Overall modeling strategy
Bootstrap, Validating and Describing the Model (H5)
Bootstrap L9
Model validation
Describing the model L10
R Multivariable Modeling/Validation/Presentation Software (H6, Alzola & Harrell 9.3-4) L11
Case study in OLS regression (H7)
Case study in data reduction and missing value imputation (H8 up until discussion of principal components) (H14.2,14.3)
Project: Understanding interrelationships of predictor variables, dealing with missing data, developing and validating a multiple regression model using least squares Assigned Due
Maximum Likelihood Estimation (H9 up until 9.3)
Binary Logistic Model (H10)
Model
Odds ratios
Special residual plots L33
Applications of general methods
Graphically presenting model L34
Case studies
Project: Develop and validate a binary logistic regression model Assigned 32 Due 35